acta biomaterialia - universidade do minho · received 9 may 2014 received in revised form 12...

Acta Biomaterialia 12 (2015) 227–241

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Bilayered silk/silk-nanoCaP scaffolds for osteochondral tissueengineering: In vitro and in vivo assessment of biological performance

http://dx.doi.org/10.1016/j.actbio.2014.10.0211742-7061/� 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: 3B’s Research Group—Biomaterials, Biodegradablesand Biomimetics, University of Minho, Headquarters of the European Institute ofExcellence on Tissue Engineering and Regenerative Medicine, AvePark, S. Cláudio deBarco, 4806-909 Taipas, Guimarães, Portugal. Tel.: +351 253 510908; fax: +351 253510909.

E-mail address: [email protected] (J.M. Oliveira).

Le-Ping Yan a,b, Joana Silva-Correia a,b, Mariana B. Oliveira a,b, Carlos Vilela a,b,c,d, Hélder Pereira a,b,e,f,Rui A. Sousa a,b, João F. Mano a,b, Ana L. Oliveira a,b,g, Joaquim M. Oliveira a,b,⇑, Rui L. Reis a,b

a 3B’s Research Group—Biomaterials, Biodegradables and Biomimetics, University of Minho, Headquarters of the European Institute of Excellence on Tissue Engineeringand Regenerative Medicine, AvePark, S. Cláudio de Barco, 4806-909 Taipas, Guimarães, Portugalb ICVS/3B’s—PT Government Associate Laboratory, Braga/Guimarães, Portugalc Life and Health Sciences Research Institute (ICVS), School of Health Sciences, University of Minho, Portugald Orthopedic Department, Centro Hospitalar do Alto Ave, Guimarães, Portugale Saúde Atlântica Sports Center—FC Porto Stadium, Minho University and Porto University Research Center, Porto, Portugalf Orthopedic Department, Centro Hospitalar Póvoa de Varzim, Vila do Conde, Portugalg CBQF—Center for Biotechnology and Fine Chemistry, School of Biotechnology, Portuguese Catholic University, Porto 4200-072, Portugal

a r t i c l e i n f o a b s t r a c t

Article history:Received 9 May 2014Received in revised form 12 September2014Accepted 15 October 2014Available online 23 October 2014

Keywords:Bilayered scaffoldSilk fibroinCalcium phosphateOsteochondral regenerationNanocomposite

Novel porous bilayered scaffolds, fully integrating a silk fibroin (SF) layer and a silk-nano calcium phos-phate (silk-nanoCaP) layer for osteochondral defect (OCD) regeneration, were developed. Homogeneousporosity distribution was achieved in the scaffolds, with calcium phosphate phase only retained in thesilk-nanoCaP layer. The scaffold presented compressive moduli of 0.4 MPa in the wet state. Rabbit bonemarrow mesenchymal stromal cells (RBMSCs) were cultured on the scaffolds, and good adhesion andproliferation were observed. The silk-nanoCaP layer showed a higher alkaline phosphatase level thanthe silk layer in osteogenic conditions. Subcutaneous implantation in rabbits demonstrated weakinflammation. In a rabbit knee critical size OCD model, the scaffolds firmly integrated into the host tissue.Histological and immunohistochemical analysis showed that collagen II positive cartilage and glycosami-noglycan regeneration presented in the silk layer, and de novo bone ingrowths and vessel formation wereobserved in the silk-nanoCaP layer. These bilayered scaffolds can therefore be promising candidates forOCD regeneration.

� 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Osteochondral defects (OCDs) are a common problem in joints[1,2], and include defects both in the articular cartilage and theunderlying subchondral bone. Cartilage defects are normally irre-versible, and would likely induce OCDs. Diseases arising from thesubchondral bone, such as osteochondritis dissecans and osteone-crosis, can also cause OCDs [3]. Osteochondral fracture constitutesan important cause of OCDs. Besides the knee, OCDs can also befound in the ankle, specifically in the talus [4,5]. OCDs will inducepersistent symptoms of pain and limited motion of the joint. Every

year, the healthcare cost for OCDs is about $95 billion in the UnitedStates alone [6].

Several techniques are currently used clinically to treat OCDs,including arthroscopic debridement, microfracture, osteochondral(OC) autograft transplantation and autologous chondrocyteimplantation [7]. These approaches are not ideal, since they arepalliative or induce donor site morbidity. OC tissue engineeringemerged as a promising alternative strategy for OCD regeneration[8–11]. It has been reported that the cartilage cannot spontane-ously repair without support from healthy subchondral bone[12]. Therefore, the rehabilitation of the subchondral bone shouldbe performed simultaneously with the reconstruction of the carti-lage layer.

The development of bioactive bilayered scaffolds for OCD regen-eration has been considered a desirable strategy [13–16]. Growthfactors have been introduced into the bilayered scaffolds toenhance cartilage repair [15,16]. Spatially controlled dual-growthfactors or gene-release systems have also been developed and both

228 L.-P. Yan et al. / Acta Biomaterialia 12 (2015) 227–241

the repair of cartilage and subchondral bone layers were observed[17,18]. On the other hand, the incorporation of osteoconductivematerials into bilayered scaffolds was able to promote the fast sub-chondral bone formation [19,20]. Some commercial bilayeredscaffolds, e.g. Trufit� and MaioRegen�, have been applied clinicallyin acellular strategies [21–23]. Kon et al. [21] showed that multi-layer collagen/nanohydroxyapatite scaffolds promoted bone andcartilage tissue restoration. Nevertheless, improvement in themechanical properties/stability of bilayered scaffold and optimiza-tion of the incorporation of bioactive factors in the scaffolds are stillbig challenges [17,24,25]. Other problems are related to achieving agood interface between the different layers [26].

Natural biopolymers have been used for tissue regeneration andpresented superior in vitro and in vivo compatibility [27,28].Among natural polymers, silk fibroin (SF) exhibits tunable mechan-ical properties [29], and thus has been finding a range ofapplications in tissue engineering [30–34]. Calcium phosphate(CaP)-based materials have been found to show outstandingosteoconductivity in bone regeneration [35,36]. Development ofpolymer/CaP composite scaffolds is a promising strategy to over-come the low elasticity of CaP [20,21]. Previously, robust SF andsilk/nano-sized CaP (silk-nanoCaP) scaffolds have been produced[37,38]. These scaffolds were able to promote in vivo new bone for-mation [39]. Based on these promising outcomes, our grouprecently developed a novel bilayered scaffold composed of a silklayer and a silk-nanoCaP layer for OCD regeneration and character-ized some preliminary properties of the scaffolds [40].

In this study, we aim to comprehensively understand the phys-icochemical properties of the bilayered scaffolds and evaluate theirbiological performance for OCD regeneration. The composition andstructure of the scaffold were evaluated by Fourier transform infra-red spectroscopy (FTIR), micro-computed tomography (micro-CT)and energy-dispersive X-ray spectrometry. Both dynamic and sta-tic mechanical properties were analyzed. Cytocompatibility wasevaluated by studying the viability and proliferation of rabbit bonemarrow mesenchymal stromal cells (RBMSCs) in the scaffolds.Osteogenic differentiation of the RBMSCs in the scaffold was alsoexamined. The bilayered scaffolds were implanted both subcutane-ously and in critical size OCDs of the rabbit knee. The regenerationof osteochondral tissues in these explants were characterized bymicro-CT, and by histological and immunohistochemical staining,respectively.

2. Materials and methods

2.1. Materials and reagents

Bombyx mori cocoons were supplied by the PortugueseAssociation of Parents and Friends of Mentally Disabled Citizens(APPACDM, Castelo Branco, Portugal). The other materials andreagents were purchased from Sigma–Aldrich (St Louis, MO, USA)unless mentioned otherwise.

2.2. Preparation of the bilayered scaffolds

The concentrated aqueous SF solution was obtained via a previ-ously reported procedure [37]. Regarding the preparation of thebilayered scaffolds, silk-nanoCaP scaffolds were first prepared[38]. Briefly, calcium chloride solution (6 mol l�1) and ammoniadibasic phosphate solution (3.6 mol l�1) of the same volume weresequentially added to the 16 wt.% SF solution, forming the nano-CaP particles in the silk solution. The amount of CaP introducedwas fixed at 16 wt.% (CaP/silk, w/w). The scaffolds were preparedby addition of sodium chloride particles into the suspension in asilicon mold. The mold was dried for 2 days, and then immersed

in distilled water overnight. During the following day, the silk-nanoCaP scaffolds were cut into pieces after removal from themolds. Each piece of scaffold was placed in the bottom of a new sil-icon mold and 300 ll of 16 wt.% silk solution was added on top ofthese scaffolds. Then, 600 mg of sodium chloride particles wereadded to the silk solution [38]. After drying and salt-leaching pro-cesses, the final scaffolds were obtained by lyophilization in afreeze drier (CRYODOS-80; Telstar, Barcelona, Spain). As controls,pure silk scaffolds and silk-nanoCaP scaffolds were also prepared.The pure silk scaffolds, the silk-nanoCaP scaffolds and thebilayered scaffolds are abbreviated as S16, SC16 and bilayered,respectively.

2.3. Physicochemical characterization of the bilayered scaffolds

2.3.1. Chemical analysis of the bilayered scaffoldsThe chemical composition and structural conformation of the

bilayered scaffolds were analyzed by a Fourier transform infraredspectroscopy (FTIR) under an attenuated total reflectance (ATR)model (IRPrestige-21, Shimadzu, Kyoto, Japan) [39]. At least threespecimens were used for each layer.

The CaP content in the silk-nanoCaP layer was evaluated bythermogravimetric analysis (TGA) [38]. The organic phase wasdegraded by heating the specimen in the TGA instrument (TGAQ500; TA Instruments, DE, USA). The Ca/P atomic ratio of the ashobtained after the TGA assay was studied by an energy-dispersiveX-ray detector (EDX). At least three specimens were used for bothassays.

2.3.2. Microstructure evaluation of the bilayered scaffoldsThe morphology of the scaffold was observed by scanning elec-

tron microscopy (SEM) (Nova NanoSEM 200; FEI, Hillsboro, OR,USA). Before the observation, the scaffolds were coated with onelayer of Au/Pd (SC502–314B) in a coater (E6700; QuorumTechnologies, East Grinstead, UK). Elemental analysis was per-formed in four zones around the interface area by an EDX setupinstalled in the SEM. Three independent areas were selected ineach zone, and each scanned area was 100 lm �100 lm.

Micro-CT was used to qualitatively and quantitatively evaluatethe porosity and the CaP distribution profile in the bilayered scaf-folds. The scanning of the scaffolds was conducted under 61 keVand 163 lA in the micro-CT (1072 scanner; SkyScan, Kontich,Belgium). Both the diameter and the height of the scaffolds were8 mm (silk layer: 3 mm in height; silk-nanoCaP layer: 5 mm inheight). The integration time was fixed at 1.3 s and the pixel reso-lution was 9.4 lm. Qualitative visualization of the 3-D morphologyand the different phases in the bilayered scaffolds were performedusing CTvox software (Skyscan). The porosity and CaP content dis-tribution profiles were processed in standardized software (CTAnalyser, version 1.5., Skyscan) [38]. Five scaffolds were used forthe qualitative and quantitative microstructure evaluation.

2.3.3. Mechanical tests of the scaffoldsThe wet status compressive test of the bilayered scaffolds was

performed in a universal testing machine (Instron 4505; Instron,Norwood, MA, USA). The diameter and the height of the scaffoldswere 6 and 5 mm, respectively (silk layer: 2 mm in height; silk-nanoCaP layer: 3 mm in height). Before the test, the samples werefirst hydrated in phosphate buffer saline solution (PBS) overnightat 37 �C. After removing the liquid with a tissue, the samples weretested using a previously reported protocol [39]. S16 and SC16were used as controls (5 mm in height, 6 mm in diameter). Foreach test, six specimens of each group were screened.

Dynamic mechanical analysis (DMA) was also conducted tostudy the viscoelastic properties of the bilayered scaffolds. Thesizes of the scaffolds were the same as for the compressive test.

L.-P. Yan et al. / Acta Biomaterialia 12 (2015) 227–241 229

The scaffolds were tested in a DMA instrument (TRITEC8000BDMA; Triton Technology, Lincolnshire, UK) [38]. Five samples ofeach group were tested.

2.4. In vitro degradation

The stability of the bilayered scaffolds was evaluated by enzy-matic degradation test. Protease XIV solution (1 mg l�1) was pre-pared by dissolving the enzyme in PBS. The initial dry weight ofthe scaffold was measured, and then the scaffolds were hydratedin PBS at 37 �C for 3 h, followed by immersion in 5 ml of proteasesolution. The scaffolds were the same size as those used for thecompressive test. The enzyme solution was changed every 24 h.The specimens were removed from the degradation solution atthe end of 0.5, 1, 2, 3, 5 and 7 days. The dry weight of the degradedspecimen was measured after drying the sample at 70 �C over-night. The weight loss ratio was obtained using the followingequation:

weight loss ratio ¼ ðmi �md;tÞðmiÞ

� �� 100% ð1Þ

where mi is the initial dry weight of the sample, and md,t is the dryweight of the degraded sample at each time point. S16 and SC16were used as controls. Five specimens per group were used for eachtime point.

2.5. In vitro cell studies

2.5.1. Isolation, expansion and seeding of the RBMSCsThe RBMSCs were isolated from male New Zealand White rab-

bits (Charles River, Senneville, Quebec, Canada). The maintenanceand usage of animals were approved by the Ethics Committee ofthe University of Minho. The 9 week old rabbits were killed byinjecting an overdose of anesthetic. The femurs were first separatedfrom the hind legs, followed by removing the epiphyseal heads andsubsequently flushing out the bone marrow plug by using alpha-minimum essential medium (a-MEM) (Gibco�; Life Technologies,Carlsbad, CA, USA). The a-MEM was supplemented with 10% fetalbovine serum, and 1% antibiotic–antimycotic liquid prepared with10,000 units ml�1 penicillin G sodium, 10,000 lg ml�1 streptomy-cin sulfate and 25 lg ml�1 amphotericin B as Fungizone� in 0.85%saline (Life Technologies, Carlsbad, CA, USA). The isolated RBMSCswere cultured in cell culture flasks at 37 �C in an incubator with5% CO2 atmosphere (MCO-18AIC (UV), Sanyo, Osaka, Japan). Themedium was changed for the first time after 4 days, and then chan-ged every 2 days until the cells reached �90% confluence. The cellswere then expanded until passage 2 before seeding in the scaffolds.All the scaffolds were sterilized with ethylene oxide (ETO).

For the cell seeding, bilayered scaffolds 6 mm in diameter and5 mm in height were used (silk layer: 2 mm in height; silk-nano-CaP layer: 3 mm in height). S16 and SC16 (6 mm in diameter and2 mm in height) were seeded with cells and used as controls forosteogenic differentiation. Before the cell seeding, the scaffoldswere hydrated in a-MEM overnight in the CO2 incubator.Afterwards, the scaffolds were removed from the medium andplaced into a 24-well suspension cell culture plate (Cell star;Greiner Bio-One, Kremsmuenster, Austria). RBMSCs of passage 2were detached from the flasks and a new cell suspension with acell density of 5 � 106 ml�1 was prepared (P3). The cells wereseeded onto the surface of the scaffolds, and then the scaffoldswith cells were kept in the CO2 incubator. After 3 h, the constructswere transferred to a new 24-well suspension culture plate andeach construct was supplemented with 2 ml of a-MEM. The cul-ture medium was refreshed every 2 or 3 days.

2.5.2. Viability, attachment, proliferation and differentiation of theRBMSCs

For the cell viability assay, 100,000 cells were seeded onto thebilayered scaffolds and cultured in basal condition. The live/deadnumbers of seeded cells were analyzed by Calcein AM and propidi-um iodide (Molecular Probes�; Life Technologies, Carlsbad, CA,USA) staining after culturing for 3 days. First, each construct waswashed with PBS, and then transferred into 1 ml PBS supple-mented with 1 lg calcein AM and 2 lg propidium iodide, for10 min. The samples were observed in a transmitted and reflectedlight microscope with apotome 2 (Axio Imager Z1 m; Zeiss, Jena,Germany) after rinsing twice wirh PBS. By using the accompanyingsoftware Zen, a Z-stack function was used to combine images atdifferent depths into one final image.

The quantitative cell viability of the constructs was screened bya 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay (MTS) using the CellTiter 96�

AQueous One Solution Cell Proliferation Assay Kit (Promega,Fitchburg, WI, USA), after culture for 1, 3 and 7 days. The MTS assayfollowed the manufacturer’s instruction. After 3 h of reaction withcells, 100 ll MTS solution was transferred into a 96-well cellculture plate and read in a microplate reader (Synergy HT;Bio-Tek, VT, USA) at 490 nm. The scaffolds without cells were usedas control.

The cells’ attachment on the scaffolds in basal condition wasobserved by SEM, after culturing for 7 days. Before the observation,the constructs were harvested from the medium and rinsed by PBS,followed by immersion in 10% formalin solution for at least 1 day.The fixed constructs were dehydrated by immersion in a series ofaqueous ethanol solutions, with a gradually increasing concentra-tion of ethanol (from 30% to 100%). The surface of the constructswere coated by Au/Pd and observed by SEM.

For the cell proliferation and osteogenic differentiation assay,200,000 cells were seeded onto the bilayered scaffolds. The follow-ing day, the constructs were cultured in basal medium (a-MEM) orosteogenesis medium, respectively. The osteogenic medium wasbased on the a-MEM, and supplemented with 10 mmol l�1 b-glyc-erophosphate, 50 lg ml�1 ascorbic acid (Wako Pure Chemicals,Tokyo, Japan) and 10�8 mol l�1 dexamethasone. The constructswere harvested after culturing for 7 and 14 days. At the end of eachtime point, the silk layer and the silk-nanoCaP layer were sepa-rated by a blade, and each part was placed into 1 ml ultrapurewater in a 1.5 ml centrifuge tube after rinsing in PBS. The silk layerscultured in basal or osteogenic conditions were abbreviated asCart.Basal or Cart.Osteo, respectively. The silk-nanoCaP layer cul-tured in basal or osteogenic conditions were abbreviated as Bone.-Basal or Bone.Osteo, respectively. S16 and SC16 were seeded with100,000 cells per scaffold. The quantification of the double-stranded DNA (dsDNA) was performed by using a Quant-ITPicoGreen dsDNA Assay Kit 2000 assay (Life Technologies,Carlsbad, CA, USA) according to the manufacturer’s instructions.The fluorescence intensities of the samples were recorded in amicroplate reader [14]. The same lysates for DNA assay were alsoused for alkaline phosphatase (ALP) activity quantification [14].The ALP activity of the samples was normalized by their corre-sponding DNA contents. The DNA contents or ALP activities ofthe bilayered scaffolds were obtained by combining the DNAcontents or ALP activities of the corresponding silk layer and silk-nanoCaP layer.

2.6. In vivo implantation of the bilayered scaffolds

In order to evaluate the in vivo biocompatibility and OCD regen-eration potential of the bilayered scaffolds, the scaffolds were sub-cutaneously implanted in the back and the knee OCD of male NewZealand White rabbits, respectively. The bilayered scaffolds 6 mm

Fig. 1. ATR-FTIR spectra of (a) the silk layer and (b) the silk-nanoCaP layer in thebilayered scaffolds. The inserted image is a backscattered SEM image of the silk-nanoCaP layer, showing the nanosized CaP particle (white domain) distribution inthe silk matrix (scale bar: 3 lm).


in diameter and 8 mm in height (silk layer: 3 mm; silk-nanoCaP:5 mm) were used for the subcutaneous implantation. Additionally,the bilayered scaffolds 5 mm in diameter and 5 mm in height (silklayer: 2 mm; silk-nanoCaP: 3 mm) were implanted in the OCD. Allthe rabbits for the in vivo studies were of 9–11 weeks old, with anaverage weight of 2.4 kg at the time of implantation. The scaffoldswere sterilized with ETO.

For the subcutaneous implantation, six bilayered scaffolds wereimplanted into three rabbits (2 pieces/rabbit). In each rabbit, twoskin incisions were made below the ears in the back (one on theleft and the other on the right), each �2 cm length. The scaffoldswere subcutaneously implanted into each pocket. After 4 weeks,the rabbits were killed by injecting an overdose of anesthetic andthe implanted scaffolds were retrieved. The explants were fixedin 10% formalin, and then dehydrated through graded ethanol,and finally embedded in paraffin. Sections were prepared by cut-ting the specimen into sections 5 lm thick using a microtome(Spencer 820, American Optical Company, NY, USA). The obtainedsections were stained with hematoxylin and eosin (H&E). Thedehydrated explants were also coated with Au/Pd and observedby SEM.

Regarding the implantation in critical size OCDs (4.5 mm indiameter and 5 mm in depth), nine bilayered scaffolds wereimplanted into three rabbits (3 pieces/rabbit). The rabbits wereanesthetized and the hair in the knee joints of the hind legs wascut. Two OCDs (4.5 mm in diameter and 5 mm in depth) were cre-ated in each femur using a Brace manual drill, one located betweenthe lateral and the medial condyle, the other in the opposite site ofthe patellar. The bilayered scaffolds were implanted into thedefects by press fitting. In each rabbit, one defect located betweenthe lateral and the medial condyle was empty and was used as con-trol. Four weeks post-operation, the rabbits were killed and theknees were excised. Three explants were fixed by 10% formalinand then immersed in paraffin after dehydration. Slides were pre-pared and H&E, Masson’s trichrome and Safranin O stainings wereperformed.

2.7. Immunohistochemical staining of the explants

Mouse anti-rabbit collagen II alpha 1 (Acris Antibodies GmbH,Herford, Germany) and biotinylated Sambucus nigra (Elderberry)bark lectin (SNA-lectin; Vector Laboratories, Burlingame, CA,USA) were used for the immunohistochemical staining of theexplants following the protocol provided by the supplier. Afterdeparaffinization and rehydration of the specimen slides, the anti-gens were retrieved in 0.01 mol l�1 sodium citrate buffer solutionat 95 �C for 20 min (pH 6.0), and then the endogenous peroxidaseswere inactivated by using 3% (v/v) hydrogen peroxide solution. Forcollagen II immunohistochemical staining, the specimen slideswere subsequently incubated overnight at 4 �C after addition ofthe primary antibodies. Then, the slides were incubated with sec-ondary antibody donkey anti-mouse Alexa Fluor 594 (MolecularProbes, Eugene, OR, USA). The nuclei of the cells in the slides werecounterstained with 40,6-diamidino-2-phynylindole (DAPI) solu-tion (1 lg ml�1 in PBS) for 10 min. Finally, the slides were mountedand observed in a fluorescence microscope. The SNA-lectin immu-nohistochemical analysis was performed according to the strepta-vidin–biotin peroxidase complex system (UltraVision LargeVolume Detection System Anti-Polyvalent, HRP; LabVisionCorporation, Fremont, CA, USA). After inactivation of the endoge-nous peroxidases and soaking in PBS, the slides were incubatedin protein blocking solution for 10 min followed by incubationwith the primary antibody SNA-lectin for 1 h and at room temper-ature. Sections were then sequentially washed with PBS andincubated with the streptavidin–peroxidase complex for 10 min.The immune reaction was visualized by 3,30-diamonobenzidine

(DAB; LabVision Corporation, Fremont, CA, USA) as a chromogen.All sections were counterstained with Gill-2 hematoxylin (Merck,Germany). For preparing the negative controls, the primary anti-body was omitted. The histological sections were then observedusing a light microscope.

2.8. Micro-CT analysis of the explants

Three explants were used for micro-CT observation in the wetstate, under 100 keV and 98 lA. The explants were loaded withParafilm� during the scanning to avoid evaporation of liquid. Theintegration time was fixed at 1.3 s and the pixel resolution was19.13 lm. The specimens were first scanned and the data sets wereprocessed as mentioned above (Section 2.3.2). In each specimen, acylindrical model region (height: 4 mm; diameter: 4 mm) wasused for the evaluation of porosity and CaP distribution. The top2 mm region in the cylindrical model region was considered as car-tilage domain in defect controls or as silk layer in defectsimplanted with scaffolds, and the next 2 mm region down wasconsidered as subchondral bone domain in defect controls or assilk-nanoCaP layer in defects implanted with scaffolds.

2.9. Statistical analysis

The data were presented by mean ± standard deviation (SD).The results were evaluated by one-way analysis of variance(ANOVA). The means of each group were compared by Tukey’s test,and P < 0.05 was considered statistically significant. Three inde-pendent experiments were performed for the cell viability assay,the proliferation and differentiation studies, and at least threesamples were analyzed for each time point in every experiment.

3. Results

3.1. Chemical composition and structural conformation of thebilayered scaffolds

The SF conformation and chemical composition in the bilayeredscaffolds were studied by ATR-FTIR. As shown in Fig. 1, the SF inboth layers displayed the same strong absorbance peaks at 1627and 1520 cm�1, which are characteristic peaks for b-sheet confor-mation [30]. It was noticed that both layers presented shoulderpeaks �1650 cm�1 corresponding to amorphous or silk-I

Table 1Microstructure and CaP content of the bilayered scaffolds analyzed by micro-CT.

Meanporosity (%)

Meaninterconnectivity (%)

Mean CaPcontent(vol.%)

Silk layer 82.02 ± 1.15 91.13 ± 2.32 0Silk-nanoCaP layer 62.27 ± 2.61 70.03 ± 4.62 9.60 ± 0.81


conformation [41]. The silk-nanoCaP layer presented a strong peakat 1031 cm�1, which is the characteristic vibration absorbance ofPO4�3 in the CaP [42]. The size of the nano-CaP particles was ana-

lyzed by back-scattered SEM. The inserted SEM image shows thatthe CaP particles were distributed evenly in the silk matrix, andwere �200 nm in size. TGA showed that the CaP mass ratio inthe silk-nanoCaP layer was �13.81 ± 0.63% (CaP/Silk, w/w), andthe Ca/P ratio of the ash was 1.65 ± 0.4%.

3.2. Microstructure and CaP distribution of the bilayered scaffolds

Fig. 2 shows a macroscopic image of the bilayered scaffolds. Itwas found that the scaffold presented macro/microporous andinterconnective structure in both layers. The two layers were wellintegrated by a continuous interface region. The pore size of themacropores in each layer was �300–700 lm, and the microporeswere located in the trabeculae of the macropores with size<50 lm (Fig. 2b). The interface region was < 500 lm thick andlocated flat between the two layers (Fig. 2b). EDX scanning fromthe silk-nanoCaP layer to the silk layer showed that calcium ionswere only limited in the silk-nanoCaP layer and the thin interfacearea (Fig. 2c). In the interface region, the intensity of the calciumion signal on the side close to the silk-nanoCaP layer was higherthan the one in the side of the silk layer.

The qualitative and quantitative distributions of the porosityand the CaP in the bilayered scaffolds were assessed by micro-CT. Table 1 demonstrates that both layers presented high porosityand interconnectivity, and the CaP was retained only in the

Fig. 2. The interface of the bilayered scaffolds. (a) Macroscopic image of thebilayered scaffolds (scale bar: 3 mm). (b) SEM image of the interface region in thebilayered scaffold (scale bar: 500 lm). Z1, Z2, Z3 and Z4 indicate different regionsfrom the silk layer to the silk-nanoCaP layer, around the interface area. (c) EDXelemental analysis of calcium ions in Z1, Z2, Z3 and Z4 regions.

silk-nanoCaP layer. The 3-D images showed two distinct phasesin the bilayered scaffold (Fig. 3a). The CaP (blue domain) residedonly in the silk-nanoCaP layer, without infiltration into the silklayer. Both layers displayed high interconnectivity and porosities(Fig. 3a). By changing the threshold, it was found that the CaP dis-tribution was homogeneous in the silk-nanoCaP layer (Fig. 3b). The2-D images of each layer also confirmed the interconnectivity andporous structure in each layer (Fig. 3c and d). The porosity distribu-tion profile revealed that the porosity was homogeneouslydistributed in each layer, and lower porosity was observed in thesilk-nanoCaP layer (Fig. 3e). The porosity showed a sharp decreasein the interface domain which was �0.5 mm in thickness. The CaPwas distributed evenly in the silk-nanoCaP layer (Fig. 3f). It wasnoticed that the CaP content decreased gradually in the thin inter-face region and there was no CaP in the silk layer (Fig. 3f).

3.3. Mechanical and degradation properties of the bilayered scaffolds

As shown in Fig. 4a, the wet state modulus of the bilayered scaf-folds was �0.4 MPa, similar to the ones of the controls. Thedynamic viscoelastic properties of the scaffolds were evaluatedby DMA. It was found that the storage modulus of the bilayeredscaffolds increased from around 0.5 to 0.8 MPa as the frequencyincreased from 0.1 to 20 Hz (Fig. 4c). In the frequency range tested,the storage modulus values of the bilayered scaffolds were similarto those of SC16 and higher than those of S16. All the three groupscaffolds demonstrated similar and low loss factor values for thetested frequencies. The loss factor (tan d) of the bilayered scaffoldsincreased slightly from around 0.17 to 0.23 when the frequencyincreased from 0.1 to 20 Hz. In this study, the enzymatic degrada-tion profiles of the scaffolds were analyzed by using protease XIV.It was found that S16 degraded faster than the bilayered scaffoldsand SC16 (Fig. 4b). In the first 12 h, the bilayered scaffolds lost�12% mass, and S16 and SC16 lost �14% and �7% mass, respec-tively. After 7 days of degradation, the bilayered scaffolds pre-sented �27% weight loss, and S16 and SC16 showed �43% and�22% weight loss, respectively.

3.4. Attachment, viability and proliferation of the RBMSCs on thebilayered scaffolds

The RBMSCs were seeded into the bilayered scaffolds. The live/dead assay showed that there were living cells attached on the sur-face of the scaffolds (Fig. 5a–c), after seeding for 3 days. The cellsdispersed evenly in the silk and silk-nanoCaP layers, presented aspreading morphology, and contacted with each other. Some cellsalso grow on the interface area. Cell attachment was also observedby SEM after culturing for 7 days in basal condition (Fig. 5d–f). Itwas found that the surface of the silk layer, the silk-nanoCaP layer,and the interface were fully covered by the cells and the extracel-lular matrix, in both basal and osteogenic conditions. The cells notonly adhered to the surface of the scaffolds, they also grew insidethe scaffolds to a depth of at least 1 mm (Fig. 5g, h).

Quantitative analysis of the cell viability was performed by MTSassay (Fig. 6a). It was observed that the MTS absorbance signifi-cantly increased during the culture period. Cell proliferation was

Fig. 3. Micro-CT analysis of the bilayered scaffolds. (a) 3-D image of the silk matrix (brown) and the CaP distribution (blue), and (b) 3-D image of the pure CaP distribution inthe bilayered scaffold (scale bar: 4 mm). (c) 2-D image of the silk layer, and (d) 2-D image of the silk-nanoCaP layer (scale bar: 1 mm). (e) Quantitative analysis of the porositydistribution, and (f) quantitative analysis of the CaP distribution in the bilayered scaffolds.

Fig. 4. Mechanical analysis and degradation profile of the bilayered scaffolds. (a) Wet status compressive modulus and (b) enzymatic degradation profile of the bilayeredscaffolds and the controls. (c) Storage modulus (E0) and (d) loss moduli (tan d) of the bilayered scaffolds and the controls obtained by DMA, tested at 37 �C in PBS.


screened by DNA content analysis. It was seen that the DNA con-tent of the bilayered scaffolds significantly increased from day 7to day 14, in both the basal and osteogenic conditions (Fig. 6b).At day 14, the DNA content of the bilayered scaffold in the basalcondition was higher than that in osteogenic media (Fig. 6b).

3.5. Osteogenic differentiation of RBMSCs in the bilayered scaffolds

The ALP activity from the cells seeded in the bilayered scaffoldsand the controls were normalized by their respective DNA content(Fig. 6c). It was found the ALP activity in all the groups increased

Fig. 5. The live/dead staining and attachment of RBMSCs in the bilayered scaffolds. (a–c) Calcein AM and propidium iodide staining of the bilayered scaffolds, after culturingRBMSCs in the scaffolds in basal condition for 3 days (scale bar: 400 lm). (d–f) SEM images of the cell attachment on the surface of the bilayered scaffold. (g, h) Longitudinalsection SEM images of the bilayered scaffolds showing the cell migration into the inner domain after culturing for 7 days in basal condition: (g) silk layer and (h) silk-nanoCaPlayer (scale bar: 500 lm).


from day 7 to day 14, in osteogenic conditions. In basal condition,the ALP activity showed no differences during the culture time. Inosteogenic conditions, the ALP activity of the silk-nanoCaP layerwas significantly higher than that of the silk layer at both testedtime points. The same trend was observed in the controls. TheALP activity of SC16 was higher than that of S16 in day 7 andday 14, when cultured in osteogenic conditions.

3.6. Subcutaneous implantation of the bilayered scaffolds

The in vivo compatibility of the bilayered scaffolds was assessedby subcutaneous implantation in rabbit. Fig. 7a showed that thebilayered scaffolds were still integrated after 4 weeks of implanta-tion. A layer of connective tissue adhered on the entire surface ofthe scaffolds, and no signs of infection or acute inflammation wereobserved. The SEM images of the explants showed that the connec-tive tissues not only tightly integrated to the implants, but alsofully filled the inner pores of the bilayered scaffolds (Fig. 7b). TheH&E staining image of the bilayered scaffolds showed that the con-nective tissues infiltrated into the pores of the scaffolds (Fig. 7c–e).There were some vessels formed inside the scaffolds (Fig. 7c). Onlya few macrophages were observed in the inner part of the scaffolds.There were also some fibroblasts presented in the silk-nanoCaPlayer.

3.7. Regeneration of rabbit knee OCDs by the bilayered scaffolds

The OC regeneration potential of the bilayered scaffolds wasstudied by implantation of these scaffolds in rabbit OCDs for4 weeks. The macroscopic images of the explants demonstratedthat the scaffolds were integrated well with the host tissue(Fig. 8a). The scaffolds implanted displayed no obvious mass loss.

There were no apparent signals of infection of the implants. Thedefect controls were not regenerated and formed a large void withapparent adjacent tissue collapse. The micro-CT analysis of theexplants illustrated that the defect filled with the bilayered scaf-fold presented less void space and more regular morphology com-pared with the defect control (Fig. 8b). Moreover, both theingrowths of the subchondral bone in the bottom domain andthe regeneration of cartilage in the surface area of the implantwere observed. From the micro-CT data, it is not easy to clearlydetermine the amount of CaP coming from the scaffold and thenew bone. However, based on the scanning parameter (high volt-age) chosen for this assay, the detected CaP may mainly come fromthe newly formed subchondral bone. The defect control showed nocartilage regeneration and little subchondral bone formation. Theporosity distribution showed that the defect control was emptyin the top region and filled by the tissues in the bottom region,while the defects with implants showed <20% porosity in theregion analyzed (Fig. 8c). The defect control showed only a verysmall amount of subchondral bone regeneration in the bottom,but the defect with the implant presented a large amount of CaPcontent in the silk-nanoCaP layer (Fig. 8d). The quantitative resultsof CaP content and porosity of different regions were presented inSupplementary Table 1. The defect controls showed much highervoid space than the defects with implants. It was observed thatthe CaP content in the silk-nanoCaP layer was �20% higher thanthe one from the silk layer.

The explants were further evaluated by H&E and Masson’s tri-chrome staining (Fig. 9). No acute inflammation was observed inany of the explants. The defects with scaffolds showed no collapseof adjacent tissues. The scaffolds presented a stable and integratedstructure, and were firmly integrated with the host tissues. In thesilk layer, the new cartilage formed and gradually spread from

Fig. 6. (a) The MTS analysis, (b) DNA content and (c) osteogenetic differentiation of the RBMSCs cultured in the bilayered scaffolds. Basal: Basal condition; Osteo: Osteogeniccondition. ⁄Indicates significant differences between the different time points. & indicates significant differences compared with DNA content from the osteogenic condition.# and % indicate significant differences compared with ALP activity from S16 group in the osteogenic condition at the 7th and 14th days, respectively. @ and § indicatesignificant differences compared with values from the silk layer in the osteogenic condition at the 7th and 14th days, respectively.


the edge to the middle of the defects (Fig. 9a,b). The top surface ofthe scaffolds showed no collapse, and matched the height of thenormal cartilage. In the bottom domain, obvious new subchondralbone growth into the silk-nanoCaP layer was observed (Fig. 9a,b).The infiltration of the subchondral bone was limited to the silk-nanoCaP layer. From the cross-sectional staining, a low level ofdeformation was observed. In addition, de novo bone infiltrationwas observed in the subchondral part, which shows a good integra-tion of the scaffolds in the host tissue (Fig. 9c,d). The defect controlshowed no regeneration of the cartilage layer and the subchondralbone (Fig. 9e,f). From the high-magnification Masson’s trichromestaining images (Fig. 10a), it could be seen that the chondrocytesin the neocartilage presented a normal round phenotype and grewinside the silk layer. The defect control only showed loose tissueformation (Fig. 10c). In the silk-nanoCaP layer, besides the obviousnew bone ingrowth, the formation of new vessels was alsodetected (Fig. 10b). Some collagen fiber formed under the loose tis-sue in the subchondral bone domain of the defect control(Fig. 10d).

The formation of glycosaminoglycan (GAG) in the defect wasstudied by Safranin O staining. It was found that the explants exhi-bit GAG tissue-containing chondrocytes that formed in the edgeand extended into the interior region of the top silk layer(Fig. 11a,b). However, the defect control did not present any posi-tive staining of cartilage matrix (Fig. 11c,d). With the collagen IIimmunohistochemical staining, it was found that the neocartilagetissue formed inside the silk layer was positively stained for colla-gen II (Fig. 12a). The defect control did not show positive stainingof collagen II (Fig. 12c). Considering the important role of angio-genesis in bone regeneration, an immunohistochemical staining

of an angiogenic marker (i.e. SNA-lectin) was performed (Fig. 13).The specificity of SNA-lectin for endothelial cells enabled observa-tion of a substantial invasion of the silk-nanoCaP layer by theendothelial cells (Fig. 13a). In the defect control (Fig. 13c), it waspossible to observe the endothelial cells mainly located in the sur-face area of the defect. The enlarged SNA-lectin staining imageshowed that abundant endothelial cells colonized the interiorregion of the silk-nanoCaP layer, and the formation of new boneand blood vessel was also identified (Fig. 13e). The defect controldid not present new bone formation and displayed much lowerendothelial cell density compared with the defect with implant(Fig. 13g). The newly formed vessels in the defect control werenot as evident as those in the defect with implants (Fig. 13e,g).

4. Discussion

OC tissue encompasses interconnected chondral and subchon-dral bone layers, which present distinct properties. Only chondro-cytes reside in the avascular chondral layer, with type II collagenand proteoglycan being the main components in the extracellularmatrix. On the other hand, many kinds of cells (e.g. osteoblasts,osteoclasts and bone marrow cells) exist in the vascularized sub-chondral bone layer which primarily consists of hydroxyapatiteand type I collagen. Therefore, in OC tissue engineering, it is feasi-ble to generate bilayered scaffolds that possess different features(e.g. mechanical, chemical or morphology properties) in each layerto satisfy the simultaneous regeneration of the chondral layer andthe subchondral bone layer. The ideal scaffolds for OCD regenera-tion should be those that can promote cartilage regeneration on

Fig. 7. Subcutaneous implantation of the bilayered scaffolds in rabbit for 4 weeks. (a) Macroscopic image of the explants after implantation for 4 weeks (scale bar: 1 cm). (b)SEM image of the explants after implantation for 4 weeks (scale bar: 1 mm), the arrow indicates the interface. (c–e) H&E staining of the silk layer, interface and silk-nanoCaPlayer in the explants after implantation for 4 weeks, respectively (scale bar: 200 lm). The arrow in (c) indicates vessels, and the arrow in (e) indicates fibroblasts.

Fig. 8. Macroscopic image and micro-CT analysis of the explants after implantation in rabbit OCD for 4 weeks. (a) Macroscopic image of the explants; the black arrowindicates the implanted scaffold, and the white arrow indicates the defect control (scale bar: 5 mm). (b) Micro-CT 3-D image of the explants; the grey arrow indicatesneocartilage, and the white arrow indicates new subchondral bone formation (scale bar: 4 mm). (c) Porosity and (d) CaP content distribution of the defect control and thedefect implanted with the bilayered scaffold.


Fig. 9. The histological analysis of the explants from rabbit OCD. (a, b) H&E staining and Masson’s trichrome staining of the longitudinal section of the explants, respectively.(c, d) H&E staining and Masson’s trichrome staining of the cross-section of the explants in the silk-nanoCaP layer, respectively. (e, f) H&E staining and Masson’s trichromestaining of the longitudinal section of the defect, respectively. The black arrow indicates neocartilage formation in the silk layer of the bilayered scaffolds, and the white arrowindicates new subchondral bone formation inside the silk-nanoCaP layer of the bilayered scaffolds. Scale bar: 1 mm.


one side and encourage fast subchondral bone repair on the other.In this study, we proposed bilayered silk/silk-nanoCaP scaffolds forOCD regeneration and evaluated their potential by in vitro andin vivo studies. These bilayered scaffolds are based on our previousworks on high-strength SF scaffolds and nanocomposite silk/nano-calcium phosphate scaffolds [37,38,43].

Compared to previous studies [37,38], the bilayered scaffoldsmaintained the chemical properties (Fig. 1 and TGA) and themicrostructure (Figs. 2 and 3) of S16 and SC16 in the chondraland subchondral layer, respectively. Tissue regeneration requiresthat scaffolds present a porous and interconnected structure, aswell as a suitable pore size and distribution [44]. It has beenreported that a pore size larger than 300 lm is favorable for cellproliferation, nutrients exchange and new tissue formation in boneregeneration; and the micropores related to the surface roughnessare good for cell attachment. Thus, the macro/microporous struc-ture in the subchondral layer may promote better cell attachment,bone tissue filtration and angiogenesis compared with the solemacroporous or microporous structure.

Based on their adequate performance for bone regeneration[35], it is believed that CaP-based materials also favor subchondralbone regeneration. However, it was also reported that CaP can

induce the hypertrophy of chondrocytes [45]. It is important toprecisely control the distribution of the CaP in the bilayered scaf-folds, avoiding the CaP migration to the layer free of CaP(addressed to cartilage). EDX (Fig. 2b,c) and micro-CT (Fig. 3a,b,f)analyses demonstrated that the introduced CaP was clearly con-fined to the silk-nanoCaP layer.

The mechanical property of the scaffolds is one of the mainissues when addressing bone and cartilage tissue regeneration[51]. Improving the mechanical properties of the scaffolds is amajor challenge for skeletal-related tissue engineering. Some stud-ies have been performed to produce silk-based scaffolds of highstrength for bone tissue engineering [52,53]. The typical Young’smodulus of articular cartilage was reported to be �1 MPa [54].Despite the different testing approach, the strength of the bilay-ered scaffolds is comparable to that of human articular cartilage(Fig. 4a). Table 2 listed the mechanical properties of previouslyreported natural polymer scaffolds. It was reported previouslythe bilayered scaffolds have a dry compressive modulus of�16 MPa [40]. In comparison with other previous studies on silk-based scaffolds, the bilayered scaffolds described herein presentsuperior compressive modulus [30,47,48,50,46]. Furthermore, themodulus of the bilayered scaffolds was much higher than those

Fig. 10. High-magnification Masson’s trichrome staining images of the explants. (a) Neocartilage formed in the top silk layer. (b) Cross-section of the explants in the silk-nanoCaP layer. ‘‘NB’’ indicates the formed new bone inside the scaffold, ‘‘S’’ indicates scaffold and the arrow indicates new vessel formation. (cm d) Cartilage and subchondralbone domains in the defect control, respectively. Scale bar: 200 lm.

Fig. 11. Safranin O staining images of the explants. (a, b) Glycosaminoglycan and proteoglycan tissues formed in the edge and the inner domain of the top silk layer,respectively. Arrow indicates newly formed tissue, ‘‘S’’ indicates scaffold. (c, d) Tissue formation in the cartilage and subchondral bone domains of the defect control,respectively. Scale bar: 200 lm.


of polysaccharide- and protein-based scaffolds (Table 2). Scaffoldsimplanted in the body undergo dynamic loading, and thus it is cru-cial to know their dynamic mechanical properties under physiolog-ical conditions. It was demonstrated that the binding strength ofthe two layers was excellent, as the bilayered scaffolds maintainedtheir integrity under high-frequency loading. The low loss factorshowed that these scaffolds were of low viscosity and highelasticity.

In an ideal tissue engineering approach the scaffolds shoulddegrade in the body after implantation. In previous studies, it

was shown that salt-leaching scaffolds gradually lost weight whenimmersed in isotonic solution for 1 year [39]. SC16 displayed aslightly higher weight loss profile as compared with S16, due tothe dissolution of CaP. When implanted, the scaffolds would con-tact with the body fluid, which is rich in enzymes. The study ofthe enzymatic degradation of scaffolds is of great importance topredict their in vivo stability. The bilayered scaffold, containingboth silk layer and silk-nanoCaP layer, presented a degradationratio in between those observed for S16 and SC16. The differencesin the degradation profiles could be attributed to the formation of

Fig. 12. Collagen II immunohistochemical staining of the explants. (a, c) Collagen II staining for the neocartilage formed in the silk layer and the defect control, respectively.(b, d) Negative control for (a, c), respectively. (a, b) Arrows indicate collagen II (red) in the regenerated cartilage, blue domain indicates the scaffold. Scale bar: 100 lm.


silk–CaP complex in the silk-nanoCaP layer. The silk and CaPformed a nanosize complex which hindered the enzymatic degra-dation process.

In order to achieve good tissue regeneration outcomes, it is nec-essary to evaluate the cytotoxicity of the scaffolds by seeding cellsonto the scaffolds and observing cell viability, attachment, prolifer-ation and differentiation. Bone marrow stromal cells are multipo-tent somatic stem cells and can be differentiated into osteoblasts,chondrocytes and adipocytes [16]. They have been studied as cellsources for OCD regeneration [16]. The outstanding performanceof the bilayered scaffolds for cell seeding was related to the intrin-sic properties of SF and CaP, as well as to the porous structure ofthe scaffolds. SF fiber has long been used as a suture for wounds[55] due to its compatibility with human tissues. SF-based bioma-terials can be proteolytically degraded and subsequently metabo-lized in vivo [55]. SF has been prepared into membranes andscaffolds for cell culture or implantation, and the outcomes weresatisfactory [29]. CaP-based materials, presenting similar chemicalproperties to the inorganic phase in bone, have thus been devel-oped into implantation materials for bone regeneration [56]. Thesalt-leaching/freeze-drying approach endowed the bilayered scaf-folds with proper hydration property [40], suitable pore size, highporosity and interconnectivity, all of which are important for sup-porting cell ingrowth, migration, proliferation and nutrienttransportation.

Interestingly, the silk-nanoCaP layer induced a higher ALP activ-ity as compared with the silk layer when the bilayered scaffoldswere cultured in osteogenic medium (Fig. 6c). ALP is an importantmarker of osteogenesis differentiation. This indicates that theincorporated CaP facilitated the RBMSCs toward osteogenic differ-entiation. The observation from S16 and SC16 also confirmed thisconclusion. This result also confirmed that the silk layer was notaffect by the silk-nanoCaP layer during the osteogenic differentia-tion of RBMSCs, and the silk-nanoCaP layer was suitable for sub-chondral bone regeneration. This data can be related to theprevious observation that only the silk-nanoCaP layer supportsapatite crystal formation when the bilayered scaffolds areimmersed in simulated body fluid [40]. Our observation is also inline with the previous study on CaP/silk hybrid scaffolds [57]. In

that study, the CaP/silk scaffolds induced higher ALP activity com-pared with the silk scaffold.

The compatibility between the bilayered scaffolds and in vivotissues was first analyzed by subcutaneous implantation. Due tothe crystalline structure of the SF, the bilayered scaffolds main-tained their integrity and kept their shapes. This result alsorevealed that the scaffolds retained good mechanical strengthin vivo, which is in good agreement with the in vitro wet compres-sive modulus analysis. These scaffolds can support tissue ingrowthand angiogenesis, present good biocompatibility in vivo, and onlyinduce minimum foreign body reaction (Fig. 7). The high porosityand interconnected structure (Figs. 2 and 3) contribute to the tis-sue ingrowth and angiogenesis of the bilayered scaffolds. Similarly,a previous study on SF-fiber-reinforced SF scaffolds showedinflammatory cells surrounded all the scaffolds at the first weekand fewer inflammatory cells at the fourth week, in a subcutaneousmice model [52].

To evaluate the potential of the bilayered scaffolds for OCDregeneration, rabbit OCD was used as a model in this study.Previously, it was reported that collagen/nanohydroxyapatite scaf-folds (MaioRegen�) would swell in contacted with blood, causingproblems with fixation [21]. In this study, the fixing of the scaffoldswas easy and the scaffolds maintained their dimensions withoutobvious swelling when in contact with body fluid during theimplantation, which is of clinical relevance. It is obvious that sig-nificant swelling of the scaffolds could cause problems duringimplantation, such as scaffold protrusion into the meniscusdomain. Besides the good stability observed for the bilayered scaf-folds, these scaffolds also promoted subchondral bone regenera-tion (Fig. 8d and Supplementary Table 1). Based on the scanningcondition used for the explants during the micro-CT analysis, theCaP content detected in the silk-nanoCaP layer mainly came fromthe newly formed subchondral bone. In a clinical trial, it wasreported that the failure in the reconstruction of patellar camefrom the slow regeneration of subchondral bone [22]. The promis-ing results in this study showed that the implantation of the bilay-ered scaffolds in a rabbit knee osteochondral critical size defect caninduce rapid subchondral bone integration and healing. The fastformation of new subchondral bone is critical to fix the implant

Fig. 13. SNA-lectin immunohistochemical staining of the explants. (a, c) SNA-lectin staining for the endothelial cells invading the silk-nanoCaP layer (indicated by arrows)and the defect control, respectively. (b, d) Negative control for (a, c), respectively. (e, g) High-magnification SNA-lectin staining images of the silk-nanoCaP layer and the defectcontrol, respectively (arrow indicated vessels). (f, h) Negative control for (e, g), respectively. ‘‘S’’ indicates scaffold, and ‘‘NB’’ indicates new bone. Scale bars: (a-d) 500 lm; (e-h) 100 lm.


in the defect site, as well as to provide mechanical support for theregeneration of cartilage layer. In the empty defect, only a smallamount of subchondral bone was formed (Fig. 8d).

The regeneration of OC tissue was further evaluated by histo-logical studies and immunohistochemistry. One of the challengesfor OCD regeneration is the formation of normal hyaline cartilage,the main components of which are collagen II and GAG [10]. Thecurrent study showed that the top silk layer was advantageous insupporting cartilage regeneration, as indicated by significant colla-gen II formation (Fig. 12a) and GAG production (Fig. 11a,b), as wellas assisting the infiltration of chondrocytes with normal

morphology (Figs. 10 and 11). Another issue adversely affectingOCD healing is rapid subchondral bone formation and vasculariza-tion within the bone [34]. It has been reported that angiogenesis isan important process during the formation of pre-vascular net-works, and endothelial cells play a unique role in angiogenesis[34]. Encouragingly, angiogenesis was clearly observed in thesilk-nanoCaP layer (Fig. 13a,e). Furthermore, the silk-nanoCaPlayer could induce new bone ingrowth and vessel formation(Figs. 10b and 13e); the latter would be helpful for new boneformation. These results have shown that the silk/silk-nanoCaPbilayered scaffolds support the rabbit knee OCD regeneration.

Table 2Compressive modulus of 3-D porous natural polymer scaffolds.

Scaffold materials Compressive modulus (kPa), dry state Compressive modulus (kPa), wet state Porosity (%)

Silka 1,300 ± 40 75 92±1.3Silkb 1,000 ± 75 <10 98 ± 1.0Silkc �40 �93Chitosanc �35 �80Collagen Id 6.31 ± 0.33 75 ± 8Hyaluronic acidd 1.33 ± 0.20 80 ± 01Gelatine 801 ± 108 97.51Gelatinf 80 ± 8 �97Silk/CaPg 16,700 ± 4,500 400 ± 102 Table 1

a Salt-leached SF scaffolds derived from 8% aqueous silk solution [30].b Salt-leached SF scaffolds derived from 17% silk solution dissolved in hexafluoroisopropanol [46,47].c Freeze-dried scaffold derived from 2% aqueous solution [48].d Freeze-dried scaffold derived 1% aqueous solution [49].e Scaffold prepared by a combination of thermally induced phase separation and porogen leaching technique [50].f Commercial gelatin scaffold: Gelfoam� [50].g Scaffolds prepared by a combination of salt-leaching/freeze-drying approach; the dry state modulus came from Ref. [40].


5. Conclusions

This study proposed novel porous bilayered scaffolds, built upby fully integrating a SF layer and a silk-nanoCaP layer, for OCDregeneration. These scaffolds presented superior mechanical prop-erties and suitable stability, as well as spatially controllable poros-ity and CaP distribution and confinement. The scaffolds supportedthe attachment, viability and proliferation of RBMSCs in vitro.Additionally, the silk-nanoCaP layer promoted better osteogenesisdifferentiation of RBMSCs in osteogenic conditions as comparedwith the silk layer. Furthermore, these scaffolds allowed tissueingrowth and induced only a very weak foreign body reactionwhen subcutaneously implanted in rabbit. When implanted in rab-bit knee critical OCDs, the bilayered scaffolds were integrated wellwith the host tissues. Moreover, the bilayered scaffolds supportedcartilage regeneration in the top silk layer, and encouraged largeamounts of subchondral bone ingrowth and angiogenesis in thebottom silk-nanoCaP layer. Although a long-term in vivo study isnecessary, the preliminary in vivo data of the bilayered scaffoldstogether with their other desirable features confirm that the silk/silk-nanoCaP bilayered scaffolds are suitable for OCD regeneration.

Conflict of interest

The authors have no conflicting interests to declare.

Acknowledgments

This study was funded by the Portuguese Foundation forScience and Technology (FCT) projects Tissue2Tissue (PTDC/CTM/105703/2008) and OsteoCart (PTDC/CTM-BPC/115977/2009), aswell as the European Union’s FP7 Programme under grantagreement no REGPOT-CT2012-316331-POLARIS. The authors aregrateful to Viviana P. Ribeiro and Teresa Oliveira for the assistancewith the immunohistochemical staining. L-P.Y. was awarded a FCTPhD scholarship (SFRH/BD/64717/2009). The FCT distinctionattributed to J.M.O. under the Investigator FCT program (IF/00423/2012) is also greatly acknowledged.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 2, 3, 5, 7–13 aredifficult to interpret in black and white. The full colour images canbe found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2014.10.021.

Appendix B. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.actbio.2014.10.021.

References

[1] Mano JF, Reis RL. Osteochondral defects: present situation and tissueengineering approaches. J Tissue Eng Regener Med 2007;1:261–73.

[2] Martin I, Miot S, Barbero A, Jakob M, Wendt D. Osteochondral tissueengineering. J Biomech 2007;40:750–65.

[3] Gomoll AH, Madry H, Knutsen G, van Dijk N, Seil R, Brittberg M, et al. Thesubchondral bone in articular cartilage repair: current problems in the surgicalmanagement. Knee Surg Sports Traumatol Arthrosc 2010;18:434–47.

[4] Weil Jr L. Biologics in foot and ankle surgery. Foot Ankle Spec 2011;4:249–52.[5] Zengerink M, Struijs PA, Tol JL, van Dijk CN. Treatment of osteochondral lesions

of the talus: a systematic review. Knee Surg Sports Traumatol Arthrosc2010;18:238–46.

[6] Bitton R. The economic burden of osteoarthritis. Am J Manag Care2009;15:S230–5.

[7] Panseri S, Russo A, Cunha C, Bondi A, Di Martino A, Patella S, et al.Osteochondral tissue engineering approaches for articular cartilage andsubchondral bone regeneration. Knee Surg Sports Traumatol Arthrosc2012;20:1182–91.

[8] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920–6.[9] Wang X, Wenk E, Zhang X, Meinel L, Vunjak-Novakovic G, Kaplan DL. Growth

factor gradients via microsphere delivery in biopolymer scaffolds forosteochondral tissue engineering. J Control Release 2009;134:81–90.

[10] Shao XX, Hutmacher DW, Ho ST, Goh JC, Lee EH. Evaluation of a hybridscaffold/cell construct in repair of high-load-bearing osteochondral defects inrabbits. Biomaterials 2006;27:1071–80.

[11] Zhang W, Chen J, Tao J, Hu C, Chen L, Zhao H, et al. The promotion ofosteochondral repair by combined intra-articular injection of parathyroidhormone-related protein and implantation of a bi-layer collagen-silk scaffold.Biomaterials 2013;34:6046–57.

[12] Minas T, Gomoll AH, Rosenberger R, Royce RO, Bryant T. Increased failure rateof autologous chondrocyte implantation after previous treatment withmarrow stimulation techniques. Am J Sports Med 2009;37:902–8.

[13] O’Shea TM, Miao X. Bilayered scaffolds for osteochondral tissue engineering.Tissue Eng Part B Rev 2008;14:447–64.

[14] Oliveira JM, Rodrigues MT, Silva SS, Malafaya PB, Gomes ME, Viegas CA, et al.Novel hydroxyapatite/chitosan bilayered scaffold for osteochondral tissue-engineering applications: scaffold design and its performance when seededwith goat bone marrow stromal cells. Biomaterials 2006;27:6123–37.

[15] Reyes R, Delgado A, Sanchez E, Fernandez A, Hernandez A, Evora C. Repair of anosteochondral defect by sustained delivery of BMP-2 or TGF-beta1 from abilayered alginate-PLGA scaffold. J Tissue Eng Regener Med 2014;8:521–33.

[16] Guo X, Park H, Liu G, Liu W, Cao Y, Tabata Y, et al. In vitro generation of anosteochondral construct using injectable hydrogel composites encapsulatingrabbit marrow mesenchymal stem cells. Biomaterials 2009;30:2741–52.

[17] Chen J, Chen H, Li P, Diao H, Zhu S, Dong L, et al. Simultaneous regeneration ofarticular cartilage and subchondral bone in vivo using MSCs induced by aspatially controlled gene delivery system in bilayered integrated scaffolds.Biomaterials 2011;32:4793–805.

[18] Kim K, Lam J, Lu S, Spicer PP, Lueckgen A, Tabata Y, et al. Osteochondral tissueregeneration using a bilayered composite hydrogel with modulating dualgrowth factor release kinetics in a rabbit model. J Control Release2013;168:166–78.


[19] Mohan N, Dormer NH, Caldwell KL, Key VH, Berkland CJ, Detamore MS.Continuous gradients of material composition and growth factors for effectiveregeneration of the osteochondral interface. Tissue Eng Part A2011;17:2845–55.

[20] Xue D, Zheng Q, Zong C, Li Q, Li H, Qian S, et al. Osteochondral repair usingporous poly(lactide-co-glycolide)/nano-hydroxyapatite hybrid scaffolds withundifferentiated mesenchymal stem cells in a rat model. J Biomed Mater Res A2010;94:259–70.

[21] Kon E, Delcogliano M, Filardo G, Busacca M, Di Martino A, Marcacci M. Novelnano-composite multilayered biomaterial for osteochondral regeneration: apilot clinical trial. Am J Sports Med 2011;39:1180–90.

[22] Joshi N, Reverte-Vinaixa M, Diaz-Ferreiro EW, Dominguez-Oronoz R. Syntheticresorbable scaffolds for the treatment of isolated patellofemoral cartilagedefects in young patients: magnetic resonance imaging and clinical evaluation.Am J Sports Med 2012;40:1289–95.

[23] Bedi A, Foo LF, Williams Iii RJ, Potter HG. The maturation of synthetic scaffoldsfor osteochondral donor sites of the knee: An MRI and T2-mapping analysis.Cartilage 2010;1:20–8.

[24] Moutos FT, Freed LE, Guilak F. A biomimetic three-dimensional wovencomposite scaffold for functional tissue engineering of cartilage. Nat Mater2007;6:162–7.

[25] Kon E, Vannini F, Buda R, Filardo G, Cavallo M, Ruffilli A, et al. How to treatosteochondritis dissecans of the knee: surgical techniques and new trendsAAOS exhibit selection. J Bone Joint Surg Am 2012;94. e1(1–8).

[26] Yang PJ, Temenoff JS. Engineering orthopedic tissue interfaces. Tissue Eng PartB Rev 2009;15:127–41.

[27] Grayson WL, Chao PH, Marolt D, Kaplan DL, Vunjak-Novakovic G. Engineeringcustom-designed osteochondral tissue grafts. Trends Biotechnol2008;26:181–9.

[28] Yan LP, Wang YJ, Ren L, Wu G, Caridade SG, Fan JB, et al. Genipin-cross-linkedcollagen/chitosan biomimetic scaffolds for articular cartilage tissueengineering applications. J Biomed Mater Res A 2010;95A:465–75.

[29] Vepari C, Kaplan DL. Silk as a biomaterial. Prog Polym Sci 2007;32:991–1007.[30] Kim UJ, Park J, Joo Kim H, Wada M, Kaplan DL. Three-dimensional

aqueous-derived biomaterial scaffolds from silk fibroin. Biomaterials2005;26:2775–85.

[31] Oliveira AL, Sun L, Kim HJ, Hu X, Rice W, Kluge J, et al. Aligned silk-based 3-Darchitectures for contact guidance in tissue engineering. Acta Biomater2012;8:1530–42.

[32] Correia C, Bhumiratana S, Yan LP, Oliveira AL, Gimble JM, Rockwood D, et al.Development of silk-based scaffolds for tissue engineering of bone fromhuman adipose-derived stem cells. Acta Biomater 2012;8:2483–92.

[33] Kundu B, Kundu SC. Osteogenesis of human stem cells in silk biomaterial forregenerative therapy. Prog Polym Sci 2010;35:1116–27.

[34] Fuchs S, Jiang X, Schmidt H, Dohle E, Ghanaati S, Orth C, et al. Dynamicprocesses involved in the pre-vascularization of silk fibroin constructs for boneregeneration using outgrowth endothelial cells. Biomaterials2009;30:1329–38.

[35] Dorozhkin SV, Epple M. Biological and medical significance of calciumphosphates. Angew Chem Int Ed 2002;41:3130–46.

[36] Salgado AJ, Coutinho OP, Reis RL. Bone tissue engineering: state of the art andfuture trends. Macromol Biosci 2004;4:743–65.

[37] Yan LP, Oliveira JM, Oliveira AL, Caridade SG, Mano JF, Reis RL. Macro/microporous silk fibroin scaffolds with potential for articular cartilage andmeniscus tissue engineering applications. Acta Biomater 2012;8:289–301.

[38] Yan LP, Silva-Correia J, Correia C, Caridade SG, Fernandes EM, Sousa RA, et al.Bioactive macro/micro porous silk fibroin/nano-sized calcium phosphatescaffolds with potential for bone-tissue-engineering applications. Nanomed(London) 2013;8:359–78.

[39] Yan LP, Salgado AJ, Oliveira JM, Oliveira AL, Reis RL. De novo bone formation onmacro/microporous silk and silk/nano-sized calcium phosphate scaffolds. JBioact Compat Polym 2013;28:439–52.

[40] Yan LP, Oliveira JM, Oliveira AL, Reis RL. Silk fibroin/nano-CaP bilayeredscaffolds for osteochondral tissue engineering. Key Eng Mater2014;587:245–8.

[41] Jin HJ, Park J, Karageorgiou V, Kim UJ, Valluzzi R, Cebe P, et al. Water-stable silkfilms with reduced b-sheet content. Adv Funct Mater 2005;15:1241–7.

[42] Oliveira JM, Silva SS, Malafaya PB, Rodrigues MT, Kotobuki N, Hirose M, et al.Macroporous hydroxyapatite scaffolds for bone tissue engineeringapplications: Physicochemical characterization and assessment of rat bonemarrow stromal cell viability. J Biomed Mater Res A 2009;91A:175–86.

[43] Oliveira AL, Sampaio SC, Sousa RA, Reis RL. Controlled mineralization ofnature-inspired silk fibroin/hydroxyapatite hybrid bioactive scaffolds for bonetissue engineering applications. In: Presented at the 20th European Conferenceon Biomaterials. Nantes, 27 September–1 October 2006.

[44] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds andosteogenesis. Biomaterials 2005;26:5474–91.

[45] Khanarian NT, Jiang J, Wan LQ, Mow VC, Lu HH. A hydrogel-mineral compositescaffold for osteochondral interface tissue engineering. Tissue Eng Part A2012;18:533–45.

[46] Nazarov R, Jin HJ, Kaplan DL. Porous 3-D scaffolds from regenerated silkfibroin. Biomacromolecules 2004;5:718–26.

[47] Marolt D, Augst A, Freed LE, Vepari C, Fajardo R, Patel N, et al. Bone andcartilage tissue constructs grown using human bone marrow stromal cells, silkscaffolds and rotating bioreactors. Biomaterials 2006;27:6138–49.

[48] Bhardwaj N, Kundu SC. Silk fibroin protein and chitosan polyelectrolytecomplex porous scaffolds for tissue engineering applications. CarbohydrPolym 2011;85:325–33.

[49] Wang TW, Spector M. Development of hyaluronic acid-based scaffolds forbrain tissue engineering. Acta Biomater 2009;5:2371–84.

[50] Liu X, Smith LA, Hu J, Ma PX. Biomimetic nanofibrous gelatin/apatitecomposite scaffolds for bone tissue engineering. Biomaterials2009;30:2252–8.

[51] Hutmacher DW. Scaffolds in tissue engineering bone and cartilage.Biomaterials 2000;21:2529–43.

[52] Mandal BB, Grinberg A, Seok Gil E, Panilaitis B, Kaplan DL. High-strength silkprotein scaffolds for bone repair. Proc Natl Acad Sci USA 2012;109:7699–704.

[53] Collins AM, Skaer NJV, Gheysens T, Knight D, Bertram C, Roach HI, et al. Bone-like resorbable silk-based scaffolds for load-bearing osteoregenerativeapplications. Adv Mater 2009;21:75–8.

[54] McMahon LA, O’Brien FJ, Prendergast PJ. Biomechanics and mechanobiology inosteochondral tissues. Regener Med 2008;3:743–59.

[55] Altman GH, Diaz F, Jakuba C, Calabro T, Horan RL, Chen JS, et al. Silk-basedbiomaterials. Biomaterials 2003;24:401–16.

[56] Oliveira JM, Kotobuki N, Tadokoro M, Hirose M, Mano JF, Reis RL, et al. Ex vivoculturing of stromal cells with dexamethasone-loadedcarboxymethylchitosan/poly(amidoamine) dendrimer nanoparticlespromotes ectopic bone formation. Bone 2010;46:1424–35.

[57] Zhang Y, Wu C, Friis T, Xiao Y. The osteogenic properties of CaP/silk compositescaffolds. Biomaterials 2010;31:2848–56.

acta biomaterialia - universidade do minho · received 9 may 2014 received in revised form 12...

Documents